[beedaddy]

Thanks to the help of DaveBlake and others, my classical music is now both well-tagged and well-processed by kodi. But there is one small (and unimportant) problem, which I nevertheless am unable to ignore.

I have the composer "Camille Saint-Saëns". But the hyphen is not correctly displayed. The hyphen is shown as a box. But this is not a sophisticated Unicode-Symbol, it's just a hyphen.

Here is a screenshot:


So, why is this? Am I doing (tagging) something wrong? Here are the tags of one of the files:

Code:

Encoded with libopus 1.1.4
User comments section follows...
        ENCODER=opusenc from opus-tools 0.1.9
        ENCODER_OPTIONS=--bitrate 160
        ENCODER=Lavf55.12.100
        CONDUCTOR=Charles Dutoit
        PERFORMER=Pascal Rogé (Klavier)
        COMPOSER=Camille Saint-Saëns
        GENRE=Klassik
        TITLE=Piano Concerto no. 1 in D major, op. 17: I. Andante. Allegro assai
        RELEASECOUNTRY=US
        TOTALDISCS=2
        LABEL=Decca Records
        TOTALTRACKS=7
        MUSICBRAINZ_ALBUMARTISTID=4137c070-15b2-4d00-a9f0-3517d02a9ba8
        MUSICBRAINZ_ALBUMARTISTID=64a7dc4d-79b0-4033-9b9a-6c64671de8b7
        MUSICBRAINZ_ALBUMARTISTID=191df02c-dd85-4197-9f50-4b3ef1486f4e
        DATE=1995-02
        DISCNUMBER=1
        TRACKTOTAL=7
        MUSICBRAINZ_RELEASETRACKID=93958ac9-516f-3458-aa97-b6d8dc775e1a
        ASIN=B00000425S
        ALBUMARTISTSORT=Saint-Saëns, Camille; Dutoit, Charles, Rogé, Pascal
        ORIGINALDATE=1986
        SCRIPT=Latn
        MUSICBRAINZ_ALBUMID=e8ad745f-2a6e-4d2d-8053-629cbce2e4b4
        RELEASESTATUS=official
        ALBUMARTIST=Camille Saint-Saëns; Charles Dutoit, Pascal Rogé
        ACOUSTID_ID=372ecda2-0a53-4f46-b934-daa1ed272fe1
        CATALOGNUMBER=00289 443 8652
        ALBUM=Piano Concertos nos. 1-5
        MUSICBRAINZ_ARTISTID=4137c070-15b2-4d00-a9f0-3517d02a9ba8
        MEDIA=CD
        ALBUMARTISTS=Camille Saint‐Saëns
        ALBUMARTISTS=Charles Dutoit
        ALBUMARTISTS=Pascal Rogé
        RELEASETYPE=album
        ORIGINALYEAR=1986
        ISRC=GBF077910021
        MUSICBRAINZ_RELEASEGROUPID=530321f2-4e5b-3192-930d-ccb78eb1d9a7
        DISCTOTAL=2
        ARTIST=Camille Saint-Saëns
        BARCODE=0028944386523
        MUSICBRAINZ_TRACKID=018918ea-205e-4f8d-8adf-8089760f5b41
        ARTISTSORT=Saint-Saëns, Camille
        ARTISTS=Camille Saint-Saëns
        TRACKNUMBER=1
        METADATA_BLOCK_PICTURE=3|image/jpeg||0x0x0|<53647 bytes of image data>
Opus stream 1:
        Pre-skip: 356
        Playback gain: 0 dB
        Channels: 2
        Original sample rate: 44100Hz
        Packet duration:   20.0ms (max),   20.0ms (avg),   20.0ms (min)
        Page duration:   1000.0ms (max),  999.5ms (avg),  640.0ms (min)
        Total data length: 15923713 bytes (overhead: 1.07%)
        Playback length: 12m:30.626s
        Average bitrate: 169.7 kb/s, w/o overhead: 167.9 kb/s
Logical stream 1 ended

[DaveBlake]

I have this artist in my library too, no problem with the hyphen (or the accents come to that).

Check the ALBUMARTISTS tag, it looks different to me, some other kind of dash maybe?

[Uatschitchun]

I strangely have the same issue with some artists, too...

One is:
http://musicbrainz.org/release/5513b36b-...tport=8000

[helta]

What font are you using?

[DaveBlake]

OK, tagging some fake files using Picard with a song from both albums I am getting some odd character issues.

It is the ALBUMARTISTS tag: viewed in both Picard and Mp3tag the dash looks the same as in the other tags, but it isn't. The other tags have char 0x2d, while ALBUMARTISTS tag has 0xe2 0x80 0x90

Picard has set this oddly, or maybe it is the script, all I can suggest is that you manually edit the ALBUMARTISTS tag using "-". I tested that and it works.

[docwra]

Yeh Musicbrainz is well known for using silly hyphen and apostrophe characters sometimes.

[Hallucyn8]

I don't know if this is related to the issue or not but I've been playing around with some of the settings in Picard and it may be worth checking the setting called "Convert Unicode punctuation characters to ASCII" under Options>Metadata, I guess it does what it says on the tin but in layman's terms standardises some of the odd punctuation symbols that seem to be in the MusicBrainz DB.

[Uatschitchun]

Nice find.
Here's a complete list of substitutions:
http://github.com/metabrainz/picard/blob...ng.py#L107

If unchecking "Convert Unicode punctuation characters to ASCII" in Picard and reloading the blink-182 album results in changes to the tags. As with a checked setting in Picard we get false characters in Kodi, not converting Unicode to ASCII is the way to go?

[Hallucyn8]

I'm generally finding I'm getting better results by having the option checked but to be fair I haven't specifically seen the issue originally reported. I'm guessing it may also depend on the regional/character settings in Kodi although this is just speculation on my part.

[Uatschitchun]

Changing the setting in Picard doen't change what kodi displays (after update & clean of db)

[beedaddy]

Thanks for the discussion. I also think that the cause is Picard. I corrected the tags with kid3, but I had to change it in the sqlite database, too.

[DaveBlake]

Yeap, it is funny data from Picard. I can see why the funny value would stay in Kodi even after you edited the tag. A rescan would say here is an artist with mbid XXX, oh yes I have that artist already, and the name not get changed. The way to avoid the hands on edit of db would be either drop the source and start again (painful scraping), or move all the music involving that artist, clean up, then scan with that music back in place.

I would just edit the db directly too.